|WARNING| fib6: fix signedness in common root build

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/171826

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-10-22

# DPDK Patch Review

## ERRORS

**Error 1: Missing `uint32_t` widening cast before shift operation**
- **Location:** Line 332 in the new code
```c
write_to_dp(tbl_ptr, ((uint64_t)tbl8_idx << 1) |
    TRIE_EXT_ENT, dp->nh_sz, 1);
```
The cast to `uint64_t` is applied to `tbl8_idx` (a signed `int` holding the result of `tbl8_alloc()`), but `tbl8_alloc()` returns `int32_t` which can have bit 31 set for valid table indices. When `tbl8_idx` is cast to `uint64_t`, if bit 31 is set, it will sign-extend to a large 64-bit value before the shift, producing an incorrect result.

**Fix:** Cast to `uint32_t` first to prevent sign extension, then to `uint64_t`:
```c
write_to_dp(tbl_ptr, ((uint64_t)(uint32_t)tbl8_idx << 1) |
    TRIE_EXT_ENT, dp->nh_sz, 1);
```

---

## WARNINGS

None identified.

---

## INFO

**Info 1: Variable declaration style**
The patch changes variable declarations from:
```c
int i, j, idx, prev_idx = 0;
```
to:
```c
uint32_t idx, prev_idx = 0;
int i, j, tbl8_idx;
```

While both the old and new styles are acceptable per DPDK guidelines (variables may be declared at the start of the block or at point of first use), the new style is clearer because it groups variables by type and purpose. This is a positive improvement but not required.

**Info 2: Correctness of the widening cast pattern**
Line 342:
```c
*tbl = get_tbl_p_by_idx(cur_tbl,
    (uint64_t)prev_idx * TRIE_TBL8_GRP_NUM_ENT, dp->nh_sz);
```
This correctly widens `prev_idx` to `uint64_t` before the multiplication to prevent overflow. This is the correct pattern per the guidelines.

---

## SUMMARY

The patch correctly addresses sign-extension issues when `get_idx()` returns values with bit 31 set, and properly widens arithmetic operations to 64 bits to prevent overflow. However, there is one remaining sign-extension bug where `tbl8_idx` (signed `int`) is cast directly to `uint64_t` before shifting, which will sign-extend if the high bit is set.

Review-Result: ERROR